Soy protein rich in isoflavones (estrogen-like compounds) has been shown to prevent bone loss in ovariectomized rats. Our short-term preliminary study results in perimenopausal women are compelling, suggesting a bone-sparing effect. These findings have prompted great interest in isoflavones as an alternative to hormone replacement therapy, yet the long-term efficacy of isoflavones on bone in humans is unknown. Our objective is to determine the three-year efficacy of isoflavone-rich soy extract in attenuating bone loss in postmenopausal women. The central hypothesis is that soy isoflavones will attenuate bone loss in early postmenopausal women by maintaining bone formation, being modulated by growth factors and isoflavone metabolism. The rationale for this research is that current hormone therapy is fraught with side effects that adversely affect women, resulting in non-compliance. This randomized double-blind placebo controlled clinical trial will examine the effects of two doses (80 or 120 mg/d) of isoflavone-rich soy extract on bone in non-osteoporotic early postmenopausal women (N=234). Specific Aims are to: 1) Determine the bone-preserving effects of isoflavones on lumbar spine bone mass as the primary outcome; 2) Relate treatment-induced changes in bone mass to changes in biochemical markers of bone turnover; 3) Identify potential mechanisms by which isoflavones prevent or modulate bone loss by measuring endogenous estrogens, sex hormone-binding globulin, insulin-like growth factor-I (IGF-I), urinary minerals, serum 25(OH)vitamin D, plasma isoflavones and their metabolites, and customary intake of isoflavone-containing soy, thus accounting for variability in response to treatment; 4) Ascertain the safety of isoflavone-rich soy extract. Caucasian women of European descent will be recruited at two sites (117 at IA, 117 at CA). Random effects repeated measures analyses will be used to: a) characterize change in bone mass as the primary outcome, b) estimate treatment-induced effects, and c) depict change in markers of bone turnover in relation to bone mass change. We will use intent-to-treat for the primary test. We will also examine potential modulators (reproductive hormones, IGF-I, plasma isoflavones) and account for other factors that affect bone, as indicated in specific aim 3. This study will provide valuable data on whether isoflavones impact bone in early postmenopausal women and help elucidate potential mechanisms, thereby contributing to our understanding of isoflavones as an alternative to traditional hormone therapy.